chimera.cpp

   1 /*
   2  *  chimera.cpp
   3  *  Mothur
   4  *
   5  *  Created by Sarah Westcott on 8/11/09.
   6  *  Copyright 2009 Schloss Lab Umass Amherst. All rights reserved.
   7  *
   8  */
   9
  10 #include "chimera.h"
  11
  12 //***************************************************************************************************************
  13 //this is a vertical soft filter
  14 string Chimera::createFilter(vector<Sequence*> seqs, float t) {
  15         try {
  16                 filterString = "";
  17                 int threshold = int (t * seqs.size());
  18 //cout << "threshhold = " << threshold << endl;
  19
  20                 vector<int> gaps;       gaps.resize(seqs[0]->getAligned().length(), 0);
  21                 vector<int> a;          a.resize(seqs[0]->getAligned().length(), 0);
  22                 vector<int> t;          t.resize(seqs[0]->getAligned().length(), 0);
  23                 vector<int> g;          g.resize(seqs[0]->getAligned().length(), 0);
  24                 vector<int> c;          c.resize(seqs[0]->getAligned().length(), 0);
  25
  26                 filterString = (string(seqs[0]->getAligned().length(), '1'));
  27
  28                 //for each sequence
  29                 for (int i = 0; i < seqs.size(); i++) {
  30
  31                         if (m->control_pressed) { return filterString; }
  32
  33                         string seqAligned = seqs[i]->getAligned();
  34
  35                         if (seqAligned.length() != filterString.length()) {  m->mothurOut(seqs[i]->getName() + " is not the same length as the template sequences. Aborting!\n");  exit(1); }
  36
  37                         for (int j = 0; j < seqAligned.length(); j++) {
  38                                 //if this spot is a gap
  39                                 if ((seqAligned[j] == '-') || (seqAligned[j] == '.'))   {       gaps[j]++;      }
  40                                 else if (toupper(seqAligned[j]) == 'A')                                 {       a[j]++;         }
  41                                 else if (toupper(seqAligned[j]) == 'T')                                 {       t[j]++;         }
  42                                 else if (toupper(seqAligned[j]) == 'G')                                 {       g[j]++;         }
  43                                 else if (toupper(seqAligned[j]) == 'C')                                 {       c[j]++;         }
  44                         }
  45                 }
  46
  47                 //zero out spot where all sequences have blanks
  48                 int numColRemoved = 0;
  49                 for(int i = 0;i < seqs[0]->getAligned().length(); i++){
  50
  51                         if (m->control_pressed) { return filterString; }
  52
  53                         if(gaps[i] == seqs.size())      {       filterString[i] = '0';  numColRemoved++;  }
  54
  55                         else if (((a[i] < threshold) && (t[i] < threshold) && (g[i] < threshold) && (c[i] < threshold))) {      filterString[i] = '0';  numColRemoved++;  }
  56                         //cout << "a = " << a[i] <<  " t = " << t[i] <<  " g = " << g[i] <<  " c = " << c[i] << endl;
  57                 }
  58
  59                 if (threshold != 0.0) {  m->mothurOut("Filter removed " + toString(numColRemoved) + " columns.");  m->mothurOutEndLine();  }
  60
  61                 return filterString;
  62         }
  63         catch(exception& e) {
  64                 m->errorOut(e, "Chimera", "createFilter");
  65                 exit(1);
  66         }
  67 }
  68 //***************************************************************************************************************
  69 map<int, int> Chimera::runFilter(Sequence* seq) {
  70         try {
  71                 map<int, int> maskMap;
  72                 string seqAligned = seq->getAligned();
  73                 string newAligned = "";
  74                 int count = 0;
  75
  76                 for (int j = 0; j < seqAligned.length(); j++) {
  77                         //if this spot is a gap
  78                         if (filterString[j] == '1') {
  79                                 newAligned += seqAligned[j];
  80                                 maskMap[count] = j;
  81                                 count++;
  82                         }
  83                 }
  84
  85                 seq->setAligned(newAligned);
  86
  87                 return maskMap;
  88         }
  89         catch(exception& e) {
  90                 m->errorOut(e, "Chimera", "runFilter");
  91                 exit(1);
  92         }
  93 }
  94 //***************************************************************************************************************
  95 vector<Sequence*> Chimera::readSeqs(string file) {
  96         try {
  97
  98                 vector<Sequence*> container;
  99                 int count = 0;
 100                 length = 0;
 101                 unaligned = false;
 102
 103                 m->mothurOut("Reading sequences from " + file + "..."); cout.flush();
 104
 105                 #ifdef USE_MPI
 106                         int pid, processors;
 107                         vector<unsigned long int> positions;
 108                         int numSeqs;
 109                         int tag = 2001;
 110
 111                         MPI_Status status;
 112                         MPI_File inMPI;
 113                         MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
 114                         MPI_Comm_size(MPI_COMM_WORLD, &processors);
 115
 116                         //char* inFileName = new char[file.length()];
 117                         //memcpy(inFileName, file.c_str(), file.length());
 118
 119                         char inFileName[1024];
 120                         strcpy(inFileName, file.c_str());
 121
 122                         MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);  //comm, filename, mode, info, filepointer
 123                         //delete inFileName;
 124
 125                         if (pid == 0) {
 126                                 positions = setFilePosFasta(file, numSeqs); //fills MPIPos, returns numSeqs
 127
 128                                 //send file positions to all processes
 129                                 for(int i = 1; i < processors; i++) {
 130                                         MPI_Send(&numSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
 131                                         MPI_Send(&positions[0], (numSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
 132                                 }
 133                         }else{
 134                                 MPI_Recv(&numSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
 135                                 positions.resize(numSeqs+1);
 136                                 MPI_Recv(&positions[0], (numSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
 137                         }
 138
 139                         //read file
 140                         for(int i=0;i<numSeqs;i++){
 141
 142                                 if (m->control_pressed) { MPI_File_close(&inMPI); return container; }
 143
 144                                 //read next sequence
 145                                 int seqlength = positions[i+1] - positions[i];
 146                                 char* buf4 = new char[seqlength];
 147
 148                                 MPI_File_read_at(inMPI, positions[i], buf4, seqlength, MPI_CHAR, &status);
 149
 150                                 string tempBuf = buf4;
 151                                 if (tempBuf.length() > seqlength) { tempBuf = tempBuf.substr(0, seqlength); }
 152                                 delete buf4;
 153
 154                                 istringstream iss (tempBuf,istringstream::in);
 155
 156                                 Sequence* current = new Sequence(iss);
 157                                 if (current->getName() != "") {
 158                                         if (count == 0) {  length = current->getAligned().length();  count++;  } //gets first seqs length
 159                                         else if (length != current->getAligned().length()) {    unaligned = true;       }
 160
 161                                         container.push_back(current);
 162                                 }
 163                         }
 164
 165                         MPI_File_close(&inMPI);
 166                         MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
 167         #else
 168
 169                 ifstream in;
 170                 openInputFile(file, in);
 171
 172                 //read in seqs and store in vector
 173                 while(!in.eof()){
 174
 175                         if (m->control_pressed) { return container; }
 176
 177                         Sequence* current = new Sequence(in);  gobble(in);
 178
 179                         if (count == 0) {  length = current->getAligned().length();  count++;  } //gets first seqs length
 180                         else if (length != current->getAligned().length()) {    unaligned = true;       }
 181
 182                         if (current->getName() != "") {  container.push_back(current);  }
 183                 }
 184                 in.close();
 185         #endif
 186
 187                 m->mothurOut("Done."); m->mothurOutEndLine();
 188
 189                 return container;
 190         }
 191         catch(exception& e) {
 192                 m->errorOut(e, "Chimera", "readSeqs");
 193                 exit(1);
 194         }
 195 }
 196 //***************************************************************************************************************
 197 void Chimera::setMask(string filename) {
 198         try {
 199
 200                 if (filename == "default") {
 201                         //default is from wigeon  236627 EU009184.1 Shigella dysenteriae str. FBD013
 202                         seqMask = ".....................................................................................................AAATTGAAGAGTTT-GA--T-CA-T-G-GCTC-AG-AT-TGAA-C-GC--TGG-C--G-GC-A-GG--C----C-T--AACACA-T-GC-A-AGT-CGA-A-CG----------G-TAA-CA-G----------------------------GAAG-A-AG----------------------------------------------------CTT-G----------------------------------------------------------------------------------CT-TCTTT----------------G-CT--G--AC--G--AG-T-GG-C-GG-A--C-------------GGG-TGAGT-A--AT-GT-C-T-G-GG---A-A--A-CT-G--C-C-TGA--TG-G------------------------------------------------------------------A-GG----GGG-AT-AA-CTA-------------------------C-T-G-----------------------GAA-A---CGG-TAG-CTAA-TA---CC-G--C-AT-A----------A--------------------C-------------------------------------GT-C-----------------------------------------------------------------------------------------------------------------------G-CA-A--------------------------------------------------------------------------------------------------------------------------------------G-A-C---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------CAAA--G-A-G-GG-----G--GA-C-CT--------------------------------------------------------------------------------------------------------------------TCG-G----------------------------------------------------------------------------------------------------------------------G----CC-TC--T---T-G--------------C----C-A---T-CG-G---AT---G-T-----G-CCC-AGA--T-GGG--A------TT--A--G-CT-A----G---TAGG-T-G-GG-G-T----AAC-GG-C-T-C-ACCT--A-GG-C-G--A-CG-A------------TCC-C-T------AG-CT-G-G-TCT-G-AG----A--GG-AT--G-AC-C-AG-CCAC-A-CTGGA--A-C-TG-A-GA-C-AC-G-G-TCCAGA-CTCC-TAC-G--G-G-A-G-GC-A-GC-A-G-TG---GG-G-A-ATA-TTGCA-C-AA-T-GG--GC-GC-A----A-G-CC-T-GA-TG-CA-GCCA-TGCC-G-CG-T---G-T-A--T--GA-A-G--A--A-G-G-CC-----TT-CG---------G-G-T-T-G-T--A---AA-G-TAC--------TT-TC-A-G--C-GGG----GA-G--G---AA-GGGA---GTAA-AG----T--T--AA-T---A----C-----CT-T-TGC-TCA-TT-GA-CG-TT-A-C-CC-G-CA-G---------AA-----------GAAGC-ACC-GG-C-TAA---C--T-CCGT--GCCA--G-C---A--GCCG---C-GG--TA-AT--AC---GG-AG-GGT-GCA-A-G-CG-TTAA-T-CGG-AA-TT-A--C-T--GGGC-GTA----AA-GCGC-AC--G-CA-G-G-C-G------------G--T-TT-G-T-T-AA----G-T-C-A---G-ATG-TG-A-AA-TC--CC-CGG-G--------------------------------------------------------------------CT-C-AA-------------------------------------------------------------------------CC-T-G-GG-AA-C----T-G-C-A-T-C--------T--GA-T-A-C-T-G-GCA--A-G-C---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------T-T-G-A-G-T-C-----T-CG--TA-G-A------------G-GG-G-GG-T----AG--AATT-CCA-G-GT--GT-A-GCG-GTGAAA-TG-CGT-AGAG-A-TC-T-GGA--GG-A-AT-A-CC-GG--T--G--GC-GAA-G--G-C---G----G--C-C-CCCTG------G-AC-GA--------------------------------------------------------------AG-A-C-T--GA--CG-----CT-CA-GG--T-G-CGA--AA-G-C--------------G-TGGG-GAG-C-A-AACA--GG-ATTA-G-ATA-C-----CC-T-G-GTA-G-T----C-CA--C-G-CCG-T-AAA--C-GATG-TC--GA-CT---------T-GG--A--G-G-TT-G-TG-C--C--------------------------------------------------------------------------------------CTT-GA--------------------------------------------------------------------------------------------------------------------------------------------------G-G-C-GT--G-G-C-T-TC-C------GG--A----GC-TAA--CG-C-G-T--T--AA-GT--C----G-ACC-GCC-T-G-GG-GAG-TA---CGG-----C-C--G-C-A-A-GGT-T--AAA-ACTC-AAA---------TGAA-TTG-ACGGG-G-G-CCCG----C-A--C-A-A-GCG-GT-G--G--AG-CA-T--GT-GGT-TT-AATT-C-G-ATG-CAAC-G-CG-A-AG-A-A-CC-TT-A-CC-TGGTC-TT-G-AC-A-T-C--------------CAC-G-G-------------A-AG-T-T-T--TC--A-GA-G-A-T--G-A-G--A-A-T-G--T-G-----CC-------------------------------------T--TC-G------------------------------------------GG----A----A---CC-GTG---A--GA---------------------------------------------------C-A-G-G-T-GCTG-CA-TGG-CT--GTC-GTC-A-GC-TC---G-TG-TT-G--TGA-AA-TGT-T-GG-G-TT-AA-GT-CCCGC-AA--------C-GAG-CGC-A-ACC-C-T-TA--TC--C-TTTG--T-T-G-C-C---AG-C-G-----G-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------TCC------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------GG---C----C-G------------G----G---A-A--CT---------------C-A-A-A-G-GA-G--AC-T-G-CCA--G-T------------------------------------G-A---TAA----------------------------------A-C-T-G--G-A-GG-A--AGG-T--GGGG-A-TGAC-GTC--AAGT-C---ATC-A-T-G-G-C-C-CTT----AC-G--AC-C-A-GG-GC-TA-CAC-ACGTG-C--TA--CAATG---G-CGCA-T-A--C-AAA-GA-GA--------------------------------------------------------------------------------------------------A-G-C-G-A--C-CTCG-C--G---------------------------------------A-GA-G-C-----------A--A-G-CG---G----------A--CCT-C------A-T-AAAGT-GC-G-T-C-G-TAG-TCC--------GGA-T-TGGAG-TC--T-GCAA-CT-C-------------------------------------------------------------------------------------------------G-ACTCC-A-T-G-AA-G-TC-GGAAT-CG-C-TA--G-TA-AT-C-G-T----GGA-TC-A-G--A------AT--GCC-AC-G-GT-G-AAT-ACGT-T-CCCGGGCCT-TGTA----CACACCG-CCC-GTC-----A---CA--CCA-TG-GG-A--G---TGG-G-TT-GC-AAA--A-GAA------G--T-AGG-TA-G-C-T-T-AA-C-C--------------------------------------------------------------TT----C-------------------------------------------------------------------------------------------------G--GG-A--GG-G--C---GC-TTA--CC--ACT-T----T-GTG-AT-TCA------------------------TG--ACT-GGGG-TG-AAG-TCGTAACAA-GGTAA-CCGT-AGGGGAA-CCTG-CGGT-TGGATCACCTCCTTA................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................";
 203                 }else if (filename == "") {  //do nothing
 204                         seqMask = "";
 205                 }else{
 206
 207         #ifdef USE_MPI
 208                         MPI_File inMPI;
 209                         MPI_Offset size;
 210                         MPI_Status status;
 211
 212                         //char* inFileName = new char[filename.length()];
 213                         //memcpy(inFileName, filename.c_str(), filename.length());
 214
 215                         char inFileName[1024];
 216                         strcpy(inFileName, filename.c_str());
 217
 218                         MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);  //comm, filename, mode, info, filepointer
 219                         MPI_File_get_size(inMPI, &size);
 220
 221                         //delete inFileName;
 222
 223                         char* buffer = new char[size];
 224                         MPI_File_read(inMPI, buffer, size, MPI_CHAR, &status);
 225
 226                         string tempBuf = buffer;
 227                         if (tempBuf.length() > size) { tempBuf = tempBuf.substr(0, size);  }
 228                         istringstream iss (tempBuf,istringstream::in);
 229
 230                         delete buffer;
 231
 232                         if (!iss.eof()) {
 233                                 Sequence temp(iss);
 234                                 seqMask = temp.getAligned();
 235                         }else {
 236                                 m->mothurOut("Problem with mask."); m->mothurOutEndLine();
 237                                 seqMask = "";
 238                         }
 239
 240                         MPI_File_close(&inMPI);
 241         #else
 242
 243                         ifstream infile;
 244                         openInputFile(filename, infile);
 245
 246                         if (!infile.eof()) {
 247                                 Sequence temp(infile);
 248                                 seqMask = temp.getAligned();
 249                         }else {
 250                                 m->mothurOut("Problem with mask."); m->mothurOutEndLine();
 251                                 seqMask = "";
 252                         }
 253                         infile.close();
 254         #endif
 255
 256                 }
 257         }
 258         catch(exception& e) {
 259                 m->errorOut(e, "Chimera", "setMask");
 260                 exit(1);
 261         }
 262 }
 263 //***************************************************************************************************************
 264 Sequence* Chimera::getSequence(string name) {
 265         try{
 266                 Sequence* temp;
 267
 268                 //look through templateSeqs til you find it
 269                 int spot = -1;
 270                 for (int i = 0; i < templateSeqs.size(); i++) {
 271                         if (name == templateSeqs[i]->getName()) {
 272                                 spot = i;
 273                                 break;
 274                         }
 275                 }
 276
 277                 if(spot == -1) { m->mothurOut("Error: Could not find sequence."); m->mothurOutEndLine(); return NULL; }
 278
 279                 temp = new Sequence(templateSeqs[spot]->getName(), templateSeqs[spot]->getAligned());
 280
 281                 return temp;
 282         }
 283         catch(exception& e) {
 284                 m->errorOut(e, "Chimera", "getSequence");
 285                 exit(1);
 286         }
 287 }
 288 //***************************************************************************************************************
 289
 290
 291
 292